| With the advent of the 5G era,mobile computing(MEC)as a technology that utilizes computing resources at the network edge is rapidly developing.However,workflow scheduling in MEC environments still faces many challenges: First,workflow scheduling for mobile devices needs to provide low-energy and low-latency service requirements,otherwise it will seriously affect user experience,and existing workflow scheduling algorithms still have room for improvement;Second,if the failure rate is not considered during the task scheduling process,temporary server failures will result in a lot of time and energy loss.To address these issues,this thesis proposes corresponding solutions.The related work and features of this thesis are as follows:1.To address the problem of unreasonable workflow task scheduling for mobile devices in MEC environments,this thesis proposes an improved random drift particle swarm workflow scheduling optimization algorithm based on energy consumption and latency awareness.The algorithm comprehensively considers the energy consumption constraints and latency sensitivity of mobile device workflow tasks,and establishes an optimization model with system latency minimization as the objective under energy consumption constraints.The algorithm first generates a task execution sequence according to the dependency relationship between workflow subtasks,ensuring the orderliness of task execution;then,it uses the relevant information of MEC nodes to calculate the execution energy consumption and time of the workflow;finally,it uses particle swarm algorithm to seek the optimal scheduling scheme,and adds crossover operation in the random drift particle update process,ensuring the diversity and globality of particles.2.To address the problem of conflict between multiple important indicators in workflow scheduling algorithms in MEC environments,this thesis proposes an improved NSGA-II workflow scheduling algorithm based on multi-objective optimization in MEC environments.In addition to generating task execution sequences,it also establishes a multi-objective optimization model,adopts an improved NSGA-II algorithm,ensures that the workflow scheduling scheme generates uniform Pareto frontiers,and solves the optimal scheduling location of workflow tasks.The improved NSGA-II-WS algorithm can improve the uniformity and globality of solutions.3.Based on spring boot,vue and other frameworks,this thesis implements a MEC workflow scheduling system aiming at delay optimization and multi-objective optimization.Through the test,the system function is complete. |
